
Ashtami Vijayan pilla :500224260
Melvin Raju :500223500


import os
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import tensorflow as tf
from tensorflow.keras import layers, models, backend as K
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import shutil
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.applications.inception_v3 import InceptionV3, preprocess_input
from tensorflow.keras.preprocessing import image
from scipy.linalg import sqrtm
from numpy import cov, trace, iscomplexobj
from numpy.random import random
from scipy.stats import entropy
from IPython.display import Image


from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive


from google.colab import drive
drive.mount('/content/drive')

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).


base_dir = '/content/drive/MyDrive/FaceExpressions/dataset'
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

for category in categories:
    print(f'{category}:')
    path = os.path.join(base_dir, category)
    print(os.listdir(path)[:5])  # List first 5 files in each category

Angry:
['295f94b25bd4406ecebe9791a9531b38c6e5ce69fd37ead8f4f9ffdb~angry.png', '4b33a2f540fd37977aca8ede305989e23e76738f89c1c16ca9a2a620~angry.jpg', '4fd88a0a743c84a624347e2c761bcd7015a929d31b40f9a36e7ee259~angry.png', 'cropped_emotions.231374~angry.png', 'cropped_emotions.158081~angry.png']
Surprise:
['cropped_emotions.100270~12fffff.png', 'cropped_emotions.100276~12fffff.png', 'cropped_emotions.100279~12fffff.png', 'cropped_emotions.100207~12fffff.png', 'cropped_emotions.100218~12fffff.png']
Sad:
['cropped_emotions.163492.png', 'cropped_emotions.164034.png', 'cropped_emotions.153406.png', 'cropped_emotions.164233.png', 'cropped_emotions.164347.png']
Ahegao:
['cropped_emotions.189154~ahegao.png', 'cropped_emotions.190427~ahegao.png', 'lol140~ahegao.png', 'lol795~ahegao.png']
Happy:
[]
Neutral:
['0b2f9d06b850023dae03f02fc6c56a6b4b3684b3b63331838edc91ddf.jpg', '0b195e0b51a44f69d06de0dbfb1893df73769c6b065c2574ff3693cff.jpg', '0b3161a4ccc3aa541df0a73f00d106ff0912c0bb9e4a0642e9bac59af.jpg', '0b0b6d01330839782219ab3d938414cf5efeec8cd79aa2cc3d9a9e95f.jpg', '0b00131d449254c09c9b2c3b295ec677940f889a3aac6a90e6982a5bf.jpg']


# Function to display images
def display_images(category, image_files):
    plt.figure(figsize=(20, 10))
    for i, img_file in enumerate(image_files):
        img_path = os.path.join(base_dir, category, img_file)
        img = mpimg.imread(img_path)
        plt.subplot(1, 5, i + 1)
        plt.imshow(img)
        plt.axis('off')
        plt.title(category)
    plt.show()

# Iterate through each category and display the first 5 images
for category in categories:
    path = os.path.join(base_dir, category)
    image_files = os.listdir(path)[:5]
    print(f'Displaying first 5 images from category: {category}')
    display_images(category, image_files)

Displaying first 5 images from category: Angry

Displaying first 5 images from category: Surprise

Displaying first 5 images from category: Sad

Displaying first 5 images from category: Ahegao

Displaying first 5 images from category: Happy

<Figure size 2000x1000 with 0 Axes>

Displaying first 5 images from category: Neutral


datagen = ImageDataGenerator(rescale=1./255)

image_size = (128, 128)  # Resize images to 128x128

train_generator = datagen.flow_from_directory(
    base_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',  # Input because we are using these images for unsupervised learning
    shuffle=True
)

Found 6793 images belonging to 6 classes.


# Create a new directory for the subset of images
subset_dir = '/content/subset_dataset'
os.makedirs(subset_dir, exist_ok=True)


# Copy 100 images from each category to the new directory
for category in categories:
    category_dir = os.path.join(base_dir, category)
    subset_category_dir = os.path.join(subset_dir, category)
    os.makedirs(subset_category_dir, exist_ok=True)

    image_files = os.listdir(category_dir)[:200]
    for img_file in image_files:
        shutil.copy(os.path.join(category_dir, img_file), subset_category_dir)


# Verify the new dataset
for category in categories:
    subset_category_dir = os.path.join(subset_dir, category)
    print(f'{category}: {len(os.listdir(subset_category_dir))} images')

Angry: 24 images
Surprise: 200 images
Sad: 200 images
Ahegao: 4 images
Happy: 0 images
Neutral: 200 images


# Define image size and latent dimension
image_size = (128, 128)
latent_dim = 2


# Data Generator
datagen = ImageDataGenerator(rescale=1./255)
train_generator = datagen.flow_from_directory(
    subset_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',
    shuffle=True
)

Found 628 images belonging to 6 classes.


# Encoder
encoder_inputs = tf.keras.Input(shape=(image_size[0], image_size[1], 3))
x = layers.Conv2D(32, 3, activation='relu', padding='same')(encoder_inputs)
x = layers.MaxPooling2D()(x)
x = layers.Conv2D(64, 3, activation='relu', padding='same')(x)
x = layers.MaxPooling2D()(x)
x = layers.Flatten()(x)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim)(x)
z_log_var = layers.Dense(latent_dim)(x)


def sampling(args):
    z_mean, z_log_var = args
    batch = K.shape(z_mean)[0]
    dim = K.int_shape(z_mean)[1]
    epsilon = K.random_normal(shape=(batch, dim))
    return z_mean + K.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling)([z_mean, z_log_var])


encoder = tf.keras.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')
encoder.summary()

Model: "encoder"
__________________________________________________________________________________________________
 Layer (type)                Output Shape                 Param #   Connected to                  
==================================================================================================
 input_1 (InputLayer)        [(None, 128, 128, 3)]        0         []                            
                                                                                                  
 conv2d (Conv2D)             (None, 128, 128, 32)         896       ['input_1[0][0]']             
                                                                                                  
 max_pooling2d (MaxPooling2  (None, 64, 64, 32)           0         ['conv2d[0][0]']              
 D)                                                                                               
                                                                                                  
 conv2d_1 (Conv2D)           (None, 64, 64, 64)           18496     ['max_pooling2d[0][0]']       
                                                                                                  
 max_pooling2d_1 (MaxPoolin  (None, 32, 32, 64)           0         ['conv2d_1[0][0]']            
 g2D)                                                                                             
                                                                                                  
 flatten (Flatten)           (None, 65536)                0         ['max_pooling2d_1[0][0]']     
                                                                                                  
 dense (Dense)               (None, 128)                  8388736   ['flatten[0][0]']             
                                                                                                  
 dense_1 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 dense_2 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 lambda (Lambda)             (None, 2)                    0         ['dense_1[0][0]',             
                                                                     'dense_2[0][0]']             
                                                                                                  
==================================================================================================
Total params: 8408644 (32.08 MB)
Trainable params: 8408644 (32.08 MB)
Non-trainable params: 0 (0.00 Byte)
__________________________________________________________________________________________________


# Decoder
latent_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(32 * 32 * 64, activation='relu')(latent_inputs)
x = layers.Reshape((32, 32, 64))(x)
x = layers.Conv2DTranspose(64, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
x = layers.Conv2DTranspose(32, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
decoder_outputs = layers.Conv2DTranspose(3, 3, activation='sigmoid', padding='same')(x)


decoder = tf.keras.Model(latent_inputs, decoder_outputs, name='decoder')
decoder.summary()

Model: "decoder"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 2)]               0         
                                                                 
 dense_3 (Dense)             (None, 65536)             196608    
                                                                 
 reshape (Reshape)           (None, 32, 32, 64)        0         
                                                                 
 conv2d_transpose (Conv2DTr  (None, 32, 32, 64)        36928     
 anspose)                                                        
                                                                 
 up_sampling2d (UpSampling2  (None, 64, 64, 64)        0         
 D)                                                              
                                                                 
 conv2d_transpose_1 (Conv2D  (None, 64, 64, 32)        18464     
 Transpose)                                                      
                                                                 
 up_sampling2d_1 (UpSamplin  (None, 128, 128, 32)      0         
 g2D)                                                            
                                                                 
 conv2d_transpose_2 (Conv2D  (None, 128, 128, 3)       867       
 Transpose)                                                      
                                                                 
=================================================================
Total params: 252867 (987.76 KB)
Trainable params: 252867 (987.76 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________


# VAE model
outputs = decoder(encoder(encoder_inputs)[2])
vae = tf.keras.Model(encoder_inputs, outputs, name='vae')


# VAE loss
reconstruction_loss = tf.keras.losses.binary_crossentropy(K.flatten(encoder_inputs), K.flatten(outputs))
reconstruction_loss *= image_size[0] * image_size[1]
kl_loss = 1 + z_log_var - K.square(z_mean) - K.exp(z_log_var)
kl_loss = K.sum(kl_loss, axis=-1)
kl_loss *= -0.5
vae_loss = K.mean(reconstruction_loss + kl_loss)
vae.add_loss(vae_loss)
vae.compile(optimizer='adam')
vae.summary()

Model: "vae"
__________________________________________________________________________________________________
 Layer (type)                Output Shape                 Param #   Connected to                  
==================================================================================================
 input_1 (InputLayer)        [(None, 128, 128, 3)]        0         []                            
                                                                                                  
 encoder (Functional)        [(None, 2),                  8408644   ['input_1[0][0]']             
                              (None, 2),                                                          
                              (None, 2)]                                                          
                                                                                                  
 decoder (Functional)        (None, 128, 128, 3)          252867    ['encoder[0][2]']             
                                                                                                  
 conv2d (Conv2D)             (None, 128, 128, 32)         896       ['input_1[0][0]']             
                                                                                                  
 max_pooling2d (MaxPooling2  (None, 64, 64, 32)           0         ['conv2d[0][0]']              
 D)                                                                                               
                                                                                                  
 conv2d_1 (Conv2D)           (None, 64, 64, 64)           18496     ['max_pooling2d[0][0]']       
                                                                                                  
 max_pooling2d_1 (MaxPoolin  (None, 32, 32, 64)           0         ['conv2d_1[0][0]']            
 g2D)                                                                                             
                                                                                                  
 flatten (Flatten)           (None, 65536)                0         ['max_pooling2d_1[0][0]']     
                                                                                                  
 dense (Dense)               (None, 128)                  8388736   ['flatten[0][0]']             
                                                                                                  
 dense_2 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 dense_1 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 tf.reshape (TFOpLambda)     (None,)                      0         ['input_1[0][0]']             
                                                                                                  
 tf.reshape_1 (TFOpLambda)   (None,)                      0         ['decoder[0][0]']             
                                                                                                  
 tf.__operators__.add (TFOp  (None, 2)                    0         ['dense_2[0][0]']             
 Lambda)                                                                                          
                                                                                                  
 tf.math.square (TFOpLambda  (None, 2)                    0         ['dense_1[0][0]']             
 )                                                                                                
                                                                                                  
 tf.cast (TFOpLambda)        (None,)                      0         ['tf.reshape[0][0]']          
                                                                                                  
 tf.convert_to_tensor (TFOp  (None,)                      0         ['tf.reshape_1[0][0]']        
 Lambda)                                                                                          
                                                                                                  
 tf.math.subtract (TFOpLamb  (None, 2)                    0         ['tf.__operators__.add[0][0]',
 da)                                                                 'tf.math.square[0][0]']      
                                                                                                  
 tf.math.exp (TFOpLambda)    (None, 2)                    0         ['dense_2[0][0]']             
                                                                                                  
 tf.keras.backend.binary_cr  (None,)                      0         ['tf.cast[0][0]',             
 ossentropy (TFOpLambda)                                             'tf.convert_to_tensor[0][0]']
                                                                                                  
 tf.math.subtract_1 (TFOpLa  (None, 2)                    0         ['tf.math.subtract[0][0]',    
 mbda)                                                               'tf.math.exp[0][0]']         
                                                                                                  
 tf.math.reduce_mean (TFOpL  ()                           0         ['tf.keras.backend.binary_cros
 ambda)                                                             sentropy[0][0]']              
                                                                                                  
 tf.math.reduce_sum (TFOpLa  (None,)                      0         ['tf.math.subtract_1[0][0]']  
 mbda)                                                                                            
                                                                                                  
 tf.math.multiply (TFOpLamb  ()                           0         ['tf.math.reduce_mean[0][0]'] 
 da)                                                                                              
                                                                                                  
 tf.math.multiply_1 (TFOpLa  (None,)                      0         ['tf.math.reduce_sum[0][0]']  
 mbda)                                                                                            
                                                                                                  
 tf.__operators__.add_1 (TF  (None,)                      0         ['tf.math.multiply[0][0]',    
 OpLambda)                                                           'tf.math.multiply_1[0][0]']  
                                                                                                  
 tf.math.reduce_mean_1 (TFO  ()                           0         ['tf.__operators__.add_1[0][0]
 pLambda)                                                           ']                            
                                                                                                  
 add_loss (AddLoss)          ()                           0         ['tf.math.reduce_mean_1[0][0]'
                                                                    ]                             
                                                                                                  
==================================================================================================
Total params: 8661511 (33.04 MB)
Trainable params: 8661511 (33.04 MB)
Non-trainable params: 0 (0.00 Byte)
__________________________________________________________________________________________________


# Training
epochs = 10
vae.fit(train_generator, epochs=epochs)

Epoch 1/10
20/20 [==============================] - 77s 3s/step - loss: 11084.1191
Epoch 2/10
20/20 [==============================] - 65s 3s/step - loss: 10786.1504
Epoch 3/10
20/20 [==============================] - 64s 3s/step - loss: 10473.9512
Epoch 4/10
20/20 [==============================] - 62s 3s/step - loss: 10304.8027
Epoch 5/10
20/20 [==============================] - 65s 3s/step - loss: 10274.2529
Epoch 6/10
20/20 [==============================] - 66s 3s/step - loss: 10219.9023
Epoch 7/10
20/20 [==============================] - 65s 3s/step - loss: 10183.4873
Epoch 8/10
20/20 [==============================] - 65s 3s/step - loss: 10173.6689
Epoch 9/10
20/20 [==============================] - 72s 4s/step - loss: 10175.3145
Epoch 10/10
20/20 [==============================] - 65s 3s/step - loss: 10160.9453

<keras.src.callbacks.History at 0x7869c07954e0>


# Load pretrained InceptionV3 model
inception_model = InceptionV3(include_top=False, pooling='avg', input_shape=(299, 299, 3))

def calculate_fid(real_images, generated_images):
    # Calculate activations
    act1 = inception_model.predict(real_images)
    act2 = inception_model.predict(generated_images)

    # Calculate mean and covariance
    mu1, sigma1 = act1.mean(axis=0), cov(act1, rowvar=False)
    mu2, sigma2 = act2.mean(axis=0), cov(act2, rowvar=False)

    # Calculate sum squared difference between means
    ssdiff = np.sum((mu1 - mu2) ** 2.0)

    # Calculate sqrt of product between covariances
    covmean = sqrtm(sigma1.dot(sigma2))

    # Check and correct imaginary numbers from sqrtm
    if iscomplexobj(covmean):
        covmean = covmean.real

    fid = ssdiff + trace(sigma1 + sigma2 - 2.0 * covmean)
    return fid

Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5
87910968/87910968 [==============================] - 2s 0us/step


# Function to preprocess images for InceptionV3
def preprocess_images(imgs):
    imgs = tf.image.resize(imgs, (299, 299))
    imgs = preprocess_input(imgs)
    return imgs


# Load Real Images
real_datagen = ImageDataGenerator(rescale=1./255)
real_generator = real_datagen.flow_from_directory(
    subset_dir,
    target_size=(299, 299),  # Resize to 299x299 for InceptionV3
    batch_size=600,  # Load all 600 real images (100 per category)
    class_mode=None,
    shuffle=True
)
real_images = next(real_generator)

Found 628 images belonging to 6 classes.


# Generate Images using VAE
def generate_images(model, latent_dim, num_images):
    z_sample = np.random.normal(size=(num_images, latent_dim))
    generated_images = model.predict(z_sample)
    return generated_images

generated_images = generate_images(decoder, latent_dim, 600)
generated_images_resized = tf.image.resize(generated_images, (299, 299))

19/19 [==============================] - 6s 327ms/step


# Preprocess images for InceptionV3
real_images_preprocessed = preprocess_input(real_images)
generated_images_preprocessed = preprocess_input(generated_images_resized)


# Calculate FID
fid = calculate_fid(real_images_preprocessed, generated_images_preprocessed)
print('FID:', fid)

19/19 [==============================] - 155s 8s/step
19/19 [==============================] - 152s 8s/step
FID: 1.3319541321978917e+53


# Ensure generated images are in the correct format (299x299x3) for InceptionV3
generated_images_resized_inception = np.array(generated_images_resized)


# Ensure images are preprocessed for InceptionV3 model
def preprocess_images(images):
    images = preprocess_input(images)
    return images


# Preprocess generated images
generated_images_preprocessed = preprocess_images(generated_images_resized_inception)

# Predict class probabilities for generated images
preds = inception_model.predict(generated_images_preprocessed)

19/19 [==============================] - 158s 8s/step


def calculate_inception_score(preds, num_splits=10):
    scores = []
    split_size = preds.shape[0] // num_splits

    for i in range(num_splits):
        split_preds = preds[i * split_size:(i + 1) * split_size]
        p_yx = np.mean(split_preds, axis=0)  # Average class probabilities across the split
        scores.append(entropy(p_yx, base=2))  # Compute entropy of the average probabilities

    inception_score = np.exp(np.mean(scores))
    return inception_score


# Calculate Inception Score
inception_score = calculate_inception_score(preds)
print('Inception Score:', inception_score)

Inception Score: 5902.863071669922


print('FID:', fid)

FID: 1.3319541321978917e+53


# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Generate one image for each category
def generate_images_per_category(model, latent_dim, num_images=1):
    generated_images = []
    for _ in range(num_images * len(categories)):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_images.append(generated_image[0])
    return np.array(generated_images)


# Generate images
generated_images = generate_images_per_category(decoder, latent_dim, num_images=1)

1/1 [==============================] - 0s 76ms/step
1/1 [==============================] - 0s 60ms/step
1/1 [==============================] - 0s 52ms/step
1/1 [==============================] - 0s 55ms/step
1/1 [==============================] - 0s 56ms/step
1/1 [==============================] - 0s 64ms/step


# Plot generated images
def plot_generated_images(images, categories):
    plt.figure(figsize=(20, 10))
    for i, img in enumerate(images):
        plt.subplot(1, len(categories), i + 1)
        plt.imshow((img * 255).astype(np.uint8))
        plt.axis('off')
        plt.title(categories[i])
    plt.show()

plot_generated_images(generated_images, categories)


# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Function to generate and save faces
def save_generated_faces(model, latent_dim, categories, save_dir='/content/generated_faces'):
    os.makedirs(save_dir, exist_ok=True)
    for i, category in enumerate(categories):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_image = (generated_image[0] * 255).astype(np.uint8)
        plt.imsave(os.path.join(save_dir, f'{category}.png'), generated_image)
save_generated_faces(decoder, latent_dim, categories)

1/1 [==============================] - 0s 59ms/step
1/1 [==============================] - 0s 51ms/step
1/1 [==============================] - 0s 59ms/step
1/1 [==============================] - 0s 52ms/step
1/1 [==============================] - 0s 50ms/step
1/1 [==============================] - 0s 61ms/step


import numpy as np
import tensorflow as tf
from tensorflow.keras import layers, models

# Example VAE definition
latent_dim = 2

# Encoder
encoder_inputs = tf.keras.Input(shape=(28, 28, 1))
x = layers.Flatten()(encoder_inputs)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim, name='z_mean')(x)
z_log_var = layers.Dense(latent_dim, name='z_log_var')(x)

# Sampling layer
def sampling(args):
    z_mean, z_log_var = args
    batch = tf.shape(z_mean)[0]
    dim = tf.shape(z_mean)[1]
    epsilon = tf.keras.backend.random_normal(shape=(batch, dim))
    return z_mean + tf.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling, output_shape=(latent_dim,), name='z')([z_mean, z_log_var])

encoder = models.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')

# Decoder
decoder_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(128, activation='relu')(decoder_inputs)
x = layers.Dense(28 * 28, activation='sigmoid')(x)
decoder_outputs = layers.Reshape((28, 28, 1))(x)

decoder = models.Model(decoder_inputs, decoder_outputs, name='decoder')

# VAE
outputs = decoder(encoder(encoder_inputs)[2])
vae = models.Model(encoder_inputs, outputs, name='vae')

# Compile and train the VAE
vae.compile(optimizer='adam', loss='binary_crossentropy')
(x_train, _), (x_test, _) = tf.keras.datasets.mnist.load_data()
x_train = np.expand_dims(x_train, -1) / 255.0
vae.fit(x_train, x_train, epochs=10, batch_size=32)

Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz
11490434/11490434 [==============================] - 0s 0us/step
Epoch 1/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.2306
Epoch 2/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.2022
Epoch 3/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1964
Epoch 4/10
1875/1875 [==============================] - 16s 9ms/step - loss: 0.1929
Epoch 5/10
1875/1875 [==============================] - 16s 9ms/step - loss: 0.1905
Epoch 6/10
1875/1875 [==============================] - 18s 10ms/step - loss: 0.1889
Epoch 7/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.1875
Epoch 8/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.1865
Epoch 9/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1856
Epoch 10/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1847

<keras.src.callbacks.History at 0x7869b3c3f460>


vae.save('vae_model.h5')

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(


encoder.save('encoder_model.h5')
decoder.save('decoder_model.h5')

WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.
WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.


from google.colab import files

# Download the VAE model
files.download('vae_model.h5')

# Download the encoder and decoder models
files.download('encoder_model.h5')
files.download('decoder_model.h5')


pip install streamlit

Collecting streamlit
  Downloading streamlit-1.36.0-py2.py3-none-any.whl (8.6 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 8.6/8.6 MB 13.1 MB/s eta 0:00:00
Requirement already satisfied: altair<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.2.2)
Requirement already satisfied: blinker<2,>=1.0.0 in /usr/lib/python3/dist-packages (from streamlit) (1.4)
Requirement already satisfied: cachetools<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (5.4.0)
Requirement already satisfied: click<9,>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.1.7)
Requirement already satisfied: numpy<3,>=1.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (1.25.2)
Requirement already satisfied: packaging<25,>=20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (24.1)
Requirement already satisfied: pandas<3,>=1.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.0.3)
Requirement already satisfied: pillow<11,>=7.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (9.4.0)
Requirement already satisfied: protobuf<6,>=3.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.20.3)
Requirement already satisfied: pyarrow>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (14.0.2)
Requirement already satisfied: requests<3,>=2.27 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.31.0)
Requirement already satisfied: rich<14,>=10.14.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (13.7.1)
Requirement already satisfied: tenacity<9,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.5.0)
Requirement already satisfied: toml<2,>=0.10.1 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.10.2)
Requirement already satisfied: typing-extensions<5,>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.12.2)
Collecting gitpython!=3.1.19,<4,>=3.0.7 (from streamlit)
  Downloading GitPython-3.1.43-py3-none-any.whl (207 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 207.3/207.3 kB 16.1 MB/s eta 0:00:00
Collecting pydeck<1,>=0.8.0b4 (from streamlit)
  Downloading pydeck-0.9.1-py2.py3-none-any.whl (6.9 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.9/6.9 MB 21.7 MB/s eta 0:00:00
Requirement already satisfied: tornado<7,>=6.0.3 in /usr/local/lib/python3.10/dist-packages (from streamlit) (6.3.3)
Collecting watchdog<5,>=2.1.5 (from streamlit)
  Downloading watchdog-4.0.1-py3-none-manylinux2014_x86_64.whl (83 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 83.0/83.0 kB 7.4 MB/s eta 0:00:00
Requirement already satisfied: entrypoints in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.4)
Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (3.1.4)
Requirement already satisfied: jsonschema>=3.0 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (4.19.2)
Requirement already satisfied: toolz in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.12.1)
Collecting gitdb<5,>=4.0.1 (from gitpython!=3.1.19,<4,>=3.0.7->streamlit)
  Downloading gitdb-4.0.11-py3-none-any.whl (62 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 62.7/62.7 kB 3.4 MB/s eta 0:00:00
Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2.8.2)
Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2023.4)
Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2024.1)
Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2.0.7)
Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2024.7.4)
Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (3.0.0)
Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (2.16.1)
Collecting smmap<6,>=3.0.1 (from gitdb<5,>=4.0.1->gitpython!=3.1.19,<4,>=3.0.7->streamlit)
  Downloading smmap-5.0.1-py3-none-any.whl (24 kB)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->altair<6,>=4.0->streamlit) (2.1.5)
Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (23.2.0)
Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (2023.12.1)
Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.35.1)
Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.19.0)
Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.10/dist-packages (from markdown-it-py>=2.2.0->rich<14,>=10.14.0->streamlit) (0.1.2)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.8.2->pandas<3,>=1.3.0->streamlit) (1.16.0)
Installing collected packages: watchdog, smmap, pydeck, gitdb, gitpython, streamlit
Successfully installed gitdb-4.0.11 gitpython-3.1.43 pydeck-0.9.1 smmap-5.0.1 streamlit-1.36.0 watchdog-4.0.1


import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io


# Load your trained VAE model
model = load_model('/content/vae_model.h5')

# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

2024-07-22 02:03:25.305 
  Warning: to view this Streamlit app on a browser, run it with the following
  command:

    streamlit run /usr/local/lib/python3.10/dist-packages/colab_kernel_launcher.py [ARGUMENTS]
2024-07-22 02:03:25.316 Session state does not function when running a script without `streamlit run`


!pip install streamlit numpy keras matplotlib

Requirement already satisfied: streamlit in /usr/local/lib/python3.10/dist-packages (1.36.0)
Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (1.25.2)
Requirement already satisfied: keras in /usr/local/lib/python3.10/dist-packages (2.15.0)
Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (3.7.1)
Requirement already satisfied: altair<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.2.2)
Requirement already satisfied: blinker<2,>=1.0.0 in /usr/lib/python3/dist-packages (from streamlit) (1.4)
Requirement already satisfied: cachetools<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (5.4.0)
Requirement already satisfied: click<9,>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.1.7)
Requirement already satisfied: packaging<25,>=20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (24.1)
Requirement already satisfied: pandas<3,>=1.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.0.3)
Requirement already satisfied: pillow<11,>=7.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (9.4.0)
Requirement already satisfied: protobuf<6,>=3.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.20.3)
Requirement already satisfied: pyarrow>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (14.0.2)
Requirement already satisfied: requests<3,>=2.27 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.31.0)
Requirement already satisfied: rich<14,>=10.14.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (13.7.1)
Requirement already satisfied: tenacity<9,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.5.0)
Requirement already satisfied: toml<2,>=0.10.1 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.10.2)
Requirement already satisfied: typing-extensions<5,>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.12.2)
Requirement already satisfied: gitpython!=3.1.19,<4,>=3.0.7 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.1.43)
Requirement already satisfied: pydeck<1,>=0.8.0b4 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.9.1)
Requirement already satisfied: tornado<7,>=6.0.3 in /usr/local/lib/python3.10/dist-packages (from streamlit) (6.3.3)
Requirement already satisfied: watchdog<5,>=2.1.5 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.0.1)
Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (1.2.1)
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (0.12.1)
Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (4.53.1)
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (1.4.5)
Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (3.1.2)
Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (2.8.2)
Requirement already satisfied: entrypoints in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.4)
Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (3.1.4)
Requirement already satisfied: jsonschema>=3.0 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (4.19.2)
Requirement already satisfied: toolz in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.12.1)
Requirement already satisfied: gitdb<5,>=4.0.1 in /usr/local/lib/python3.10/dist-packages (from gitpython!=3.1.19,<4,>=3.0.7->streamlit) (4.0.11)
Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2023.4)
Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2024.1)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)
Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2.0.7)
Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2024.7.4)
Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (3.0.0)
Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (2.16.1)
Requirement already satisfied: smmap<6,>=3.0.1 in /usr/local/lib/python3.10/dist-packages (from gitdb<5,>=4.0.1->gitpython!=3.1.19,<4,>=3.0.7->streamlit) (5.0.1)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->altair<6,>=4.0->streamlit) (2.1.5)
Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (23.2.0)
Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (2023.12.1)
Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.35.1)
Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.19.0)
Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.10/dist-packages (from markdown-it-py>=2.2.0->rich<14,>=10.14.0->streamlit) (0.1.2)


# Write the Streamlit app to a file
%%writefile app.py
import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io

Writing app.py


# Load your trained VAE model
model = load_model('/content/vae_model.h5')


# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

    # Generate the image
    generated_image = model.predict(input_vector)

    # Scale the pixel values to [0, 255]
    generated_image = (generated_image * 127.5 + 127.5).astype(np.uint8)
    generated_image = generated_image[0]  # Remove the batch dimension

    # Display the image
    st.image(generated_image, use_column_width=True)


!pip install pyngrok

Collecting pyngrok
  Downloading pyngrok-7.2.0-py3-none-any.whl (22 kB)
Requirement already satisfied: PyYAML>=5.1 in /usr/local/lib/python3.10/dist-packages (from pyngrok) (6.0.1)
Installing collected packages: pyngrok
Successfully installed pyngrok-7.2.0


from pyngrok import ngrok

# Run Streamlit in the background
!streamlit run app.py &

# Set up ngrok to tunnel to the Streamlit app
public_url = ngrok.connect(port='8501')
print(f"Streamlit app is live at: {public_url}")

Collecting usage statistics. To deactivate, set browser.gatherUsageStats to false.


  You can now view your Streamlit app in your browser.

  Local URL: http://localhost:8501
  Network URL: http://172.28.0.12:8501
  External URL: http://34.42.123.165:8501

  Stopping...

---------------------------------------------------------------------------
KeyboardInterrupt                         Traceback (most recent call last)
<ipython-input-52-ff30c2567270> in <cell line: 7>()
      5 
      6 # Set up ngrok to tunnel to the Streamlit app
----> 7 public_url = ngrok.connect(port='8501')
      8 print(f"Streamlit app is live at: {public_url}")

/usr/local/lib/python3.10/dist-packages/pyngrok/ngrok.py in connect(addr, proto, name, pyngrok_config, **options)
    314             options.pop("auth")
    315 
--> 316     api_url = get_ngrok_process(pyngrok_config).api_url
    317 
    318     logger.debug(f"Creating tunnel with options: {options}")

/usr/local/lib/python3.10/dist-packages/pyngrok/ngrok.py in get_ngrok_process(pyngrok_config)
    152         pyngrok_config = conf.get_default()
    153 
--> 154     install_ngrok(pyngrok_config)
    155 
    156     return process.get_process(pyngrok_config)

/usr/local/lib/python3.10/dist-packages/pyngrok/ngrok.py in install_ngrok(pyngrok_config)
     97 
     98     if not os.path.exists(pyngrok_config.ngrok_path):
---> 99         installer.install_ngrok(pyngrok_config.ngrok_path, ngrok_version=pyngrok_config.ngrok_version)
    100 
    101     config_path = conf.get_config_path(pyngrok_config)

/usr/local/lib/python3.10/dist-packages/pyngrok/installer.py in install_ngrok(ngrok_path, ngrok_version, **kwargs)
    122         download_path = _download_file(url, **kwargs)
    123 
--> 124         _install_ngrok_zip(ngrok_path, download_path)
    125     except Exception as e:
    126         raise PyngrokNgrokInstallError(f"An error occurred while downloading ngrok from {url}: {e}")

/usr/local/lib/python3.10/dist-packages/pyngrok/installer.py in _install_ngrok_zip(ngrok_path, zip_path)
    139     with zipfile.ZipFile(zip_path, "r") as zip_ref:
    140         logger.debug(f"Extracting ngrok binary from {zip_path} to {ngrok_path} ...")
--> 141         zip_ref.extractall(os.path.dirname(ngrok_path))
    142 
    143     os.chmod(ngrok_path, int("777", 8))

/usr/lib/python3.10/zipfile.py in extractall(self, path, members, pwd)
   1657 
   1658         for zipinfo in members:
-> 1659             self._extract_member(zipinfo, path, pwd)
   1660 
   1661     @classmethod

/usr/lib/python3.10/zipfile.py in _extract_member(self, member, targetpath, pwd)
   1712         with self.open(member, pwd=pwd) as source, \
   1713              open(targetpath, "wb") as target:
-> 1714             shutil.copyfileobj(source, target)
   1715 
   1716         return targetpath

/usr/lib/python3.10/shutil.py in copyfileobj(fsrc, fdst, length)
    196         if not buf:
    197             break
--> 198         fdst_write(buf)
    199 
    200 def _samefile(src, dst):

KeyboardInterrupt:


# Install nbconvert if not already installed
!pip install nbconvert

# Convert the notebook to HTML (Replace 'your_notebook.ipynb' with your actual notebook name)
!jupyter nbconvert --to html '/content/Vae.ipynb'

# List the files in the current directory to confirm the HTML file creation
!ls /content

# Optional: Display the HTML file in Colab
from IPython.display import HTML
with open('/content/Vae.html', 'r') as f:
    html_data = f.read()
HTML(html_data)

Requirement already satisfied: nbconvert in /usr/local/lib/python3.10/dist-packages (6.5.4)
Requirement already satisfied: lxml in /usr/local/lib/python3.10/dist-packages (from nbconvert) (4.9.4)
Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (4.12.3)
Requirement already satisfied: bleach in /usr/local/lib/python3.10/dist-packages (from nbconvert) (6.1.0)
Requirement already satisfied: defusedxml in /usr/local/lib/python3.10/dist-packages (from nbconvert) (0.7.1)
Requirement already satisfied: entrypoints>=0.2.2 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (0.4)
Requirement already satisfied: jinja2>=3.0 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (3.1.4)
Requirement already satisfied: jupyter-core>=4.7 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (5.7.2)
Requirement already satisfied: jupyterlab-pygments in /usr/local/lib/python3.10/dist-packages (from nbconvert) (0.3.0)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (2.1.5)
Requirement already satisfied: mistune<2,>=0.8.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (0.8.4)
Requirement already satisfied: nbclient>=0.5.0 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (0.10.0)
Requirement already satisfied: nbformat>=5.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (5.10.4)
Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from nbconvert) (24.1)
Requirement already satisfied: pandocfilters>=1.4.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (1.5.1)
Requirement already satisfied: pygments>=2.4.1 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (2.16.1)
Requirement already satisfied: tinycss2 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (1.3.0)
Requirement already satisfied: traitlets>=5.0 in /usr/local/lib/python3.10/dist-packages (from nbconvert) (5.7.1)
Requirement already satisfied: platformdirs>=2.5 in /usr/local/lib/python3.10/dist-packages (from jupyter-core>=4.7->nbconvert) (4.2.2)
Requirement already satisfied: jupyter-client>=6.1.12 in /usr/local/lib/python3.10/dist-packages (from nbclient>=0.5.0->nbconvert) (6.1.12)
Requirement already satisfied: fastjsonschema>=2.15 in /usr/local/lib/python3.10/dist-packages (from nbformat>=5.1->nbconvert) (2.20.0)
Requirement already satisfied: jsonschema>=2.6 in /usr/local/lib/python3.10/dist-packages (from nbformat>=5.1->nbconvert) (4.19.2)
Requirement already satisfied: soupsieve>1.2 in /usr/local/lib/python3.10/dist-packages (from beautifulsoup4->nbconvert) (2.5)
Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python3.10/dist-packages (from bleach->nbconvert) (1.16.0)
Requirement already satisfied: webencodings in /usr/local/lib/python3.10/dist-packages (from bleach->nbconvert) (0.5.1)
Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat>=5.1->nbconvert) (23.2.0)
Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat>=5.1->nbconvert) (2023.12.1)
Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat>=5.1->nbconvert) (0.35.1)
Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=2.6->nbformat>=5.1->nbconvert) (0.19.0)
Requirement already satisfied: pyzmq>=13 in /usr/local/lib/python3.10/dist-packages (from jupyter-client>=6.1.12->nbclient>=0.5.0->nbconvert) (24.0.1)
Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.10/dist-packages (from jupyter-client>=6.1.12->nbclient>=0.5.0->nbconvert) (2.8.2)
Requirement already satisfied: tornado>=4.1 in /usr/local/lib/python3.10/dist-packages (from jupyter-client>=6.1.12->nbclient>=0.5.0->nbconvert) (6.3.3)
[NbConvertApp] Converting notebook /content/Vae.ipynb to html
[NbConvertApp] ERROR | Notebook JSON is invalid: Additional properties are not allowed ('metadata' was unexpected)

Failed validating 'additionalProperties' in stream:

On instance['cells'][66]['outputs'][0]:
{'metadata': {'tags': None},
 'name': 'stdout',
 'output_type': 'stream',
 'text': '\n'
         'Collecting usage statistics. To deactivate, set browser.gatherU...'}
[NbConvertApp] Writing 5408784 bytes to /content/Vae.html
app.py		  drive		    generated_faces  subset_dataset  Vae.ipynb
decoder_model.h5  encoder_model.h5  sample_data      Vae.html	     vae_model.h5

import os
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import tensorflow as tf
from tensorflow.keras import layers, models, backend as K
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import shutil
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.applications.inception_v3 import InceptionV3, preprocess_input
from tensorflow.keras.preprocessing import image
from scipy.linalg import sqrtm
from numpy import cov, trace, iscomplexobj
from numpy.random import random
from scipy.stats import entropy
from IPython.display import Image

from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive

from google.colab import drive
drive.mount('/content/drive')

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).

base_dir = '/content/drive/MyDrive/FaceExpressions/dataset'
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

for category in categories:
    print(f'{category}:')
    path = os.path.join(base_dir, category)
    print(os.listdir(path)[:5])  # List first 5 files in each category

Angry:
['295f94b25bd4406ecebe9791a9531b38c6e5ce69fd37ead8f4f9ffdb~angry.png', '4b33a2f540fd37977aca8ede305989e23e76738f89c1c16ca9a2a620~angry.jpg', '4fd88a0a743c84a624347e2c761bcd7015a929d31b40f9a36e7ee259~angry.png', 'cropped_emotions.231374~angry.png', 'cropped_emotions.158081~angry.png']
Surprise:
['cropped_emotions.100270~12fffff.png', 'cropped_emotions.100276~12fffff.png', 'cropped_emotions.100279~12fffff.png', 'cropped_emotions.100207~12fffff.png', 'cropped_emotions.100218~12fffff.png']
Sad:
['cropped_emotions.163492.png', 'cropped_emotions.164034.png', 'cropped_emotions.153406.png', 'cropped_emotions.164233.png', 'cropped_emotions.164347.png']
Ahegao:
['cropped_emotions.189154~ahegao.png', 'cropped_emotions.190427~ahegao.png', 'lol140~ahegao.png', 'lol795~ahegao.png']
Happy:
[]
Neutral:
['0b2f9d06b850023dae03f02fc6c56a6b4b3684b3b63331838edc91ddf.jpg', '0b195e0b51a44f69d06de0dbfb1893df73769c6b065c2574ff3693cff.jpg', '0b3161a4ccc3aa541df0a73f00d106ff0912c0bb9e4a0642e9bac59af.jpg', '0b0b6d01330839782219ab3d938414cf5efeec8cd79aa2cc3d9a9e95f.jpg', '0b00131d449254c09c9b2c3b295ec677940f889a3aac6a90e6982a5bf.jpg']

# Function to display images
def display_images(category, image_files):
    plt.figure(figsize=(20, 10))
    for i, img_file in enumerate(image_files):
        img_path = os.path.join(base_dir, category, img_file)
        img = mpimg.imread(img_path)
        plt.subplot(1, 5, i + 1)
        plt.imshow(img)
        plt.axis('off')
        plt.title(category)
    plt.show()

# Iterate through each category and display the first 5 images
for category in categories:
    path = os.path.join(base_dir, category)
    image_files = os.listdir(path)[:5]
    print(f'Displaying first 5 images from category: {category}')
    display_images(category, image_files)


import os
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import tensorflow as tf
from tensorflow.keras import layers, models, backend as K
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import shutil
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.applications.inception_v3 import InceptionV3, preprocess_input
from tensorflow.keras.preprocessing import image
from scipy.linalg import sqrtm
from numpy import cov, trace, iscomplexobj
from numpy.random import random
from scipy.stats import entropy
from IPython.display import Image

import os
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import tensorflow as tf
from tensorflow.keras import layers, models, backend as K
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import shutil
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.applications.inception_v3 import InceptionV3, preprocess_input
from tensorflow.keras.preprocessing import image
from scipy.linalg import sqrtm
from numpy import cov, trace, iscomplexobj
from numpy.random import random
from scipy.stats import entropy
from IPython.display import Image


from google.colab import drive
drive.mount('/content/drive')

from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive


from google.colab import drive
drive.mount('/content/drive')

from google.colab import drive
drive.mount('/content/drive')

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).


base_dir = '/content/drive/MyDrive/FaceExpressions/dataset'
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

for category in categories:
    print(f'{category}:')
    path = os.path.join(base_dir, category)
    print(os.listdir(path)[:5])  # List first 5 files in each category

base_dir = '/content/drive/MyDrive/FaceExpressions/dataset'
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

for category in categories:
    print(f'{category}:')
    path = os.path.join(base_dir, category)
    print(os.listdir(path)[:5])  # List first 5 files in each category

Angry:
['295f94b25bd4406ecebe9791a9531b38c6e5ce69fd37ead8f4f9ffdb~angry.png', '4b33a2f540fd37977aca8ede305989e23e76738f89c1c16ca9a2a620~angry.jpg', '4fd88a0a743c84a624347e2c761bcd7015a929d31b40f9a36e7ee259~angry.png', 'cropped_emotions.231374~angry.png', 'cropped_emotions.158081~angry.png']
Surprise:
['cropped_emotions.100270~12fffff.png', 'cropped_emotions.100276~12fffff.png', 'cropped_emotions.100279~12fffff.png', 'cropped_emotions.100207~12fffff.png', 'cropped_emotions.100218~12fffff.png']
Sad:
['cropped_emotions.163492.png', 'cropped_emotions.164034.png', 'cropped_emotions.153406.png', 'cropped_emotions.164233.png', 'cropped_emotions.164347.png']
Ahegao:
['cropped_emotions.189154~ahegao.png', 'cropped_emotions.190427~ahegao.png', 'lol140~ahegao.png', 'lol795~ahegao.png']
Happy:
[]
Neutral:
['0b2f9d06b850023dae03f02fc6c56a6b4b3684b3b63331838edc91ddf.jpg', '0b195e0b51a44f69d06de0dbfb1893df73769c6b065c2574ff3693cff.jpg', '0b3161a4ccc3aa541df0a73f00d106ff0912c0bb9e4a0642e9bac59af.jpg', '0b0b6d01330839782219ab3d938414cf5efeec8cd79aa2cc3d9a9e95f.jpg', '0b00131d449254c09c9b2c3b295ec677940f889a3aac6a90e6982a5bf.jpg']


# Function to display images
def display_images(category, image_files):
    plt.figure(figsize=(20, 10))
    for i, img_file in enumerate(image_files):
        img_path = os.path.join(base_dir, category, img_file)
        img = mpimg.imread(img_path)
        plt.subplot(1, 5, i + 1)
        plt.imshow(img)
        plt.axis('off')
        plt.title(category)
    plt.show()

# Iterate through each category and display the first 5 images
for category in categories:
    path = os.path.join(base_dir, category)
    image_files = os.listdir(path)[:5]
    print(f'Displaying first 5 images from category: {category}')
    display_images(category, image_files)

# Function to display images
def display_images(category, image_files):
    plt.figure(figsize=(20, 10))
    for i, img_file in enumerate(image_files):
        img_path = os.path.join(base_dir, category, img_file)
        img = mpimg.imread(img_path)
        plt.subplot(1, 5, i + 1)
        plt.imshow(img)
        plt.axis('off')
        plt.title(category)
    plt.show()

# Iterate through each category and display the first 5 images
for category in categories:
    path = os.path.join(base_dir, category)
    image_files = os.listdir(path)[:5]
    print(f'Displaying first 5 images from category: {category}')
    display_images(category, image_files)

Displaying first 5 images from category: Angry

Displaying first 5 images from category: Surprise

Displaying first 5 images from category: Sad

Displaying first 5 images from category: Ahegao

Displaying first 5 images from category: Happy

<Figure size 2000x1000 with 0 Axes>

Displaying first 5 images from category: Neutral


datagen = ImageDataGenerator(rescale=1./255)

image_size = (128, 128)  # Resize images to 128x128

train_generator = datagen.flow_from_directory(
    base_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',  # Input because we are using these images for unsupervised learning
    shuffle=True
)

datagen = ImageDataGenerator(rescale=1./255)

image_size = (128, 128)  # Resize images to 128x128

train_generator = datagen.flow_from_directory(
    base_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',  # Input because we are using these images for unsupervised learning
    shuffle=True
)

Found 6793 images belonging to 6 classes.


# Create a new directory for the subset of images
subset_dir = '/content/subset_dataset'
os.makedirs(subset_dir, exist_ok=True)

# Create a new directory for the subset of images
subset_dir = '/content/subset_dataset'
os.makedirs(subset_dir, exist_ok=True)


# Copy 100 images from each category to the new directory
for category in categories:
    category_dir = os.path.join(base_dir, category)
    subset_category_dir = os.path.join(subset_dir, category)
    os.makedirs(subset_category_dir, exist_ok=True)

    image_files = os.listdir(category_dir)[:200]
    for img_file in image_files:
        shutil.copy(os.path.join(category_dir, img_file), subset_category_dir)

# Copy 100 images from each category to the new directory
for category in categories:
    category_dir = os.path.join(base_dir, category)
    subset_category_dir = os.path.join(subset_dir, category)
    os.makedirs(subset_category_dir, exist_ok=True)

    image_files = os.listdir(category_dir)[:200]
    for img_file in image_files:
        shutil.copy(os.path.join(category_dir, img_file), subset_category_dir)


# Verify the new dataset
for category in categories:
    subset_category_dir = os.path.join(subset_dir, category)
    print(f'{category}: {len(os.listdir(subset_category_dir))} images')

# Verify the new dataset
for category in categories:
    subset_category_dir = os.path.join(subset_dir, category)
    print(f'{category}: {len(os.listdir(subset_category_dir))} images')

Angry: 24 images
Surprise: 200 images
Sad: 200 images
Ahegao: 4 images
Happy: 0 images
Neutral: 200 images


# Define image size and latent dimension
image_size = (128, 128)
latent_dim = 2

# Define image size and latent dimension
image_size = (128, 128)
latent_dim = 2


# Data Generator
datagen = ImageDataGenerator(rescale=1./255)
train_generator = datagen.flow_from_directory(
    subset_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',
    shuffle=True
)

# Data Generator
datagen = ImageDataGenerator(rescale=1./255)
train_generator = datagen.flow_from_directory(
    subset_dir,
    target_size=image_size,
    batch_size=32,
    class_mode='input',
    shuffle=True
)

Found 628 images belonging to 6 classes.


# Encoder
encoder_inputs = tf.keras.Input(shape=(image_size[0], image_size[1], 3))
x = layers.Conv2D(32, 3, activation='relu', padding='same')(encoder_inputs)
x = layers.MaxPooling2D()(x)
x = layers.Conv2D(64, 3, activation='relu', padding='same')(x)
x = layers.MaxPooling2D()(x)
x = layers.Flatten()(x)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim)(x)
z_log_var = layers.Dense(latent_dim)(x)

# Encoder
encoder_inputs = tf.keras.Input(shape=(image_size[0], image_size[1], 3))
x = layers.Conv2D(32, 3, activation='relu', padding='same')(encoder_inputs)
x = layers.MaxPooling2D()(x)
x = layers.Conv2D(64, 3, activation='relu', padding='same')(x)
x = layers.MaxPooling2D()(x)
x = layers.Flatten()(x)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim)(x)
z_log_var = layers.Dense(latent_dim)(x)


def sampling(args):
    z_mean, z_log_var = args
    batch = K.shape(z_mean)[0]
    dim = K.int_shape(z_mean)[1]
    epsilon = K.random_normal(shape=(batch, dim))
    return z_mean + K.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling)([z_mean, z_log_var])

def sampling(args):
    z_mean, z_log_var = args
    batch = K.shape(z_mean)[0]
    dim = K.int_shape(z_mean)[1]
    epsilon = K.random_normal(shape=(batch, dim))
    return z_mean + K.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling)([z_mean, z_log_var])


encoder = tf.keras.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')
encoder.summary()

encoder = tf.keras.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')
encoder.summary()

Model: "encoder"
__________________________________________________________________________________________________
 Layer (type)                Output Shape                 Param #   Connected to                  
==================================================================================================
 input_1 (InputLayer)        [(None, 128, 128, 3)]        0         []                            
                                                                                                  
 conv2d (Conv2D)             (None, 128, 128, 32)         896       ['input_1[0][0]']             
                                                                                                  
 max_pooling2d (MaxPooling2  (None, 64, 64, 32)           0         ['conv2d[0][0]']              
 D)                                                                                               
                                                                                                  
 conv2d_1 (Conv2D)           (None, 64, 64, 64)           18496     ['max_pooling2d[0][0]']       
                                                                                                  
 max_pooling2d_1 (MaxPoolin  (None, 32, 32, 64)           0         ['conv2d_1[0][0]']            
 g2D)                                                                                             
                                                                                                  
 flatten (Flatten)           (None, 65536)                0         ['max_pooling2d_1[0][0]']     
                                                                                                  
 dense (Dense)               (None, 128)                  8388736   ['flatten[0][0]']             
                                                                                                  
 dense_1 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 dense_2 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 lambda (Lambda)             (None, 2)                    0         ['dense_1[0][0]',             
                                                                     'dense_2[0][0]']             
                                                                                                  
==================================================================================================
Total params: 8408644 (32.08 MB)
Trainable params: 8408644 (32.08 MB)
Non-trainable params: 0 (0.00 Byte)
__________________________________________________________________________________________________


# Decoder
latent_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(32 * 32 * 64, activation='relu')(latent_inputs)
x = layers.Reshape((32, 32, 64))(x)
x = layers.Conv2DTranspose(64, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
x = layers.Conv2DTranspose(32, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
decoder_outputs = layers.Conv2DTranspose(3, 3, activation='sigmoid', padding='same')(x)

# Decoder
latent_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(32 * 32 * 64, activation='relu')(latent_inputs)
x = layers.Reshape((32, 32, 64))(x)
x = layers.Conv2DTranspose(64, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
x = layers.Conv2DTranspose(32, 3, activation='relu', padding='same')(x)
x = layers.UpSampling2D()(x)
decoder_outputs = layers.Conv2DTranspose(3, 3, activation='sigmoid', padding='same')(x)


decoder = tf.keras.Model(latent_inputs, decoder_outputs, name='decoder')
decoder.summary()

decoder = tf.keras.Model(latent_inputs, decoder_outputs, name='decoder')
decoder.summary()

Model: "decoder"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 2)]               0         
                                                                 
 dense_3 (Dense)             (None, 65536)             196608    
                                                                 
 reshape (Reshape)           (None, 32, 32, 64)        0         
                                                                 
 conv2d_transpose (Conv2DTr  (None, 32, 32, 64)        36928     
 anspose)                                                        
                                                                 
 up_sampling2d (UpSampling2  (None, 64, 64, 64)        0         
 D)                                                              
                                                                 
 conv2d_transpose_1 (Conv2D  (None, 64, 64, 32)        18464     
 Transpose)                                                      
                                                                 
 up_sampling2d_1 (UpSamplin  (None, 128, 128, 32)      0         
 g2D)                                                            
                                                                 
 conv2d_transpose_2 (Conv2D  (None, 128, 128, 3)       867       
 Transpose)                                                      
                                                                 
=================================================================
Total params: 252867 (987.76 KB)
Trainable params: 252867 (987.76 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________


# VAE model
outputs = decoder(encoder(encoder_inputs)[2])
vae = tf.keras.Model(encoder_inputs, outputs, name='vae')

# VAE model
outputs = decoder(encoder(encoder_inputs)[2])
vae = tf.keras.Model(encoder_inputs, outputs, name='vae')


# VAE loss
reconstruction_loss = tf.keras.losses.binary_crossentropy(K.flatten(encoder_inputs), K.flatten(outputs))
reconstruction_loss *= image_size[0] * image_size[1]
kl_loss = 1 + z_log_var - K.square(z_mean) - K.exp(z_log_var)
kl_loss = K.sum(kl_loss, axis=-1)
kl_loss *= -0.5
vae_loss = K.mean(reconstruction_loss + kl_loss)
vae.add_loss(vae_loss)
vae.compile(optimizer='adam')
vae.summary()

# VAE loss
reconstruction_loss = tf.keras.losses.binary_crossentropy(K.flatten(encoder_inputs), K.flatten(outputs))
reconstruction_loss *= image_size[0] * image_size[1]
kl_loss = 1 + z_log_var - K.square(z_mean) - K.exp(z_log_var)
kl_loss = K.sum(kl_loss, axis=-1)
kl_loss *= -0.5
vae_loss = K.mean(reconstruction_loss + kl_loss)
vae.add_loss(vae_loss)
vae.compile(optimizer='adam')
vae.summary()

Model: "vae"
__________________________________________________________________________________________________
 Layer (type)                Output Shape                 Param #   Connected to                  
==================================================================================================
 input_1 (InputLayer)        [(None, 128, 128, 3)]        0         []                            
                                                                                                  
 encoder (Functional)        [(None, 2),                  8408644   ['input_1[0][0]']             
                              (None, 2),                                                          
                              (None, 2)]                                                          
                                                                                                  
 decoder (Functional)        (None, 128, 128, 3)          252867    ['encoder[0][2]']             
                                                                                                  
 conv2d (Conv2D)             (None, 128, 128, 32)         896       ['input_1[0][0]']             
                                                                                                  
 max_pooling2d (MaxPooling2  (None, 64, 64, 32)           0         ['conv2d[0][0]']              
 D)                                                                                               
                                                                                                  
 conv2d_1 (Conv2D)           (None, 64, 64, 64)           18496     ['max_pooling2d[0][0]']       
                                                                                                  
 max_pooling2d_1 (MaxPoolin  (None, 32, 32, 64)           0         ['conv2d_1[0][0]']            
 g2D)                                                                                             
                                                                                                  
 flatten (Flatten)           (None, 65536)                0         ['max_pooling2d_1[0][0]']     
                                                                                                  
 dense (Dense)               (None, 128)                  8388736   ['flatten[0][0]']             
                                                                                                  
 dense_2 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 dense_1 (Dense)             (None, 2)                    258       ['dense[0][0]']               
                                                                                                  
 tf.reshape (TFOpLambda)     (None,)                      0         ['input_1[0][0]']             
                                                                                                  
 tf.reshape_1 (TFOpLambda)   (None,)                      0         ['decoder[0][0]']             
                                                                                                  
 tf.__operators__.add (TFOp  (None, 2)                    0         ['dense_2[0][0]']             
 Lambda)                                                                                          
                                                                                                  
 tf.math.square (TFOpLambda  (None, 2)                    0         ['dense_1[0][0]']             
 )                                                                                                
                                                                                                  
 tf.cast (TFOpLambda)        (None,)                      0         ['tf.reshape[0][0]']          
                                                                                                  
 tf.convert_to_tensor (TFOp  (None,)                      0         ['tf.reshape_1[0][0]']        
 Lambda)                                                                                          
                                                                                                  
 tf.math.subtract (TFOpLamb  (None, 2)                    0         ['tf.__operators__.add[0][0]',
 da)                                                                 'tf.math.square[0][0]']      
                                                                                                  
 tf.math.exp (TFOpLambda)    (None, 2)                    0         ['dense_2[0][0]']             
                                                                                                  
 tf.keras.backend.binary_cr  (None,)                      0         ['tf.cast[0][0]',             
 ossentropy (TFOpLambda)                                             'tf.convert_to_tensor[0][0]']
                                                                                                  
 tf.math.subtract_1 (TFOpLa  (None, 2)                    0         ['tf.math.subtract[0][0]',    
 mbda)                                                               'tf.math.exp[0][0]']         
                                                                                                  
 tf.math.reduce_mean (TFOpL  ()                           0         ['tf.keras.backend.binary_cros
 ambda)                                                             sentropy[0][0]']              
                                                                                                  
 tf.math.reduce_sum (TFOpLa  (None,)                      0         ['tf.math.subtract_1[0][0]']  
 mbda)                                                                                            
                                                                                                  
 tf.math.multiply (TFOpLamb  ()                           0         ['tf.math.reduce_mean[0][0]'] 
 da)                                                                                              
                                                                                                  
 tf.math.multiply_1 (TFOpLa  (None,)                      0         ['tf.math.reduce_sum[0][0]']  
 mbda)                                                                                            
                                                                                                  
 tf.__operators__.add_1 (TF  (None,)                      0         ['tf.math.multiply[0][0]',    
 OpLambda)                                                           'tf.math.multiply_1[0][0]']  
                                                                                                  
 tf.math.reduce_mean_1 (TFO  ()                           0         ['tf.__operators__.add_1[0][0]
 pLambda)                                                           ']                            
                                                                                                  
 add_loss (AddLoss)          ()                           0         ['tf.math.reduce_mean_1[0][0]'
                                                                    ]                             
                                                                                                  
==================================================================================================
Total params: 8661511 (33.04 MB)
Trainable params: 8661511 (33.04 MB)
Non-trainable params: 0 (0.00 Byte)
__________________________________________________________________________________________________


# Training
epochs = 10
vae.fit(train_generator, epochs=epochs)

# Training
epochs = 10
vae.fit(train_generator, epochs=epochs)

Epoch 1/10
20/20 [==============================] - 77s 3s/step - loss: 11084.1191
Epoch 2/10
20/20 [==============================] - 65s 3s/step - loss: 10786.1504
Epoch 3/10
20/20 [==============================] - 64s 3s/step - loss: 10473.9512
Epoch 4/10
20/20 [==============================] - 62s 3s/step - loss: 10304.8027
Epoch 5/10
20/20 [==============================] - 65s 3s/step - loss: 10274.2529
Epoch 6/10
20/20 [==============================] - 66s 3s/step - loss: 10219.9023
Epoch 7/10
20/20 [==============================] - 65s 3s/step - loss: 10183.4873
Epoch 8/10
20/20 [==============================] - 65s 3s/step - loss: 10173.6689
Epoch 9/10
20/20 [==============================] - 72s 4s/step - loss: 10175.3145
Epoch 10/10
20/20 [==============================] - 65s 3s/step - loss: 10160.9453

<keras.src.callbacks.History at 0x7869c07954e0>


# Load pretrained InceptionV3 model
inception_model = InceptionV3(include_top=False, pooling='avg', input_shape=(299, 299, 3))

def calculate_fid(real_images, generated_images):
    # Calculate activations
    act1 = inception_model.predict(real_images)
    act2 = inception_model.predict(generated_images)

    # Calculate mean and covariance
    mu1, sigma1 = act1.mean(axis=0), cov(act1, rowvar=False)
    mu2, sigma2 = act2.mean(axis=0), cov(act2, rowvar=False)

    # Calculate sum squared difference between means
    ssdiff = np.sum((mu1 - mu2) ** 2.0)

    # Calculate sqrt of product between covariances
    covmean = sqrtm(sigma1.dot(sigma2))

    # Check and correct imaginary numbers from sqrtm
    if iscomplexobj(covmean):
        covmean = covmean.real

    fid = ssdiff + trace(sigma1 + sigma2 - 2.0 * covmean)
    return fid

# Load pretrained InceptionV3 model
inception_model = InceptionV3(include_top=False, pooling='avg', input_shape=(299, 299, 3))

def calculate_fid(real_images, generated_images):
    # Calculate activations
    act1 = inception_model.predict(real_images)
    act2 = inception_model.predict(generated_images)

    # Calculate mean and covariance
    mu1, sigma1 = act1.mean(axis=0), cov(act1, rowvar=False)
    mu2, sigma2 = act2.mean(axis=0), cov(act2, rowvar=False)

    # Calculate sum squared difference between means
    ssdiff = np.sum((mu1 - mu2) ** 2.0)

    # Calculate sqrt of product between covariances
    covmean = sqrtm(sigma1.dot(sigma2))

    # Check and correct imaginary numbers from sqrtm
    if iscomplexobj(covmean):
        covmean = covmean.real

    fid = ssdiff + trace(sigma1 + sigma2 - 2.0 * covmean)
    return fid

Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5
87910968/87910968 [==============================] - 2s 0us/step


# Function to preprocess images for InceptionV3
def preprocess_images(imgs):
    imgs = tf.image.resize(imgs, (299, 299))
    imgs = preprocess_input(imgs)
    return imgs

# Function to preprocess images for InceptionV3
def preprocess_images(imgs):
    imgs = tf.image.resize(imgs, (299, 299))
    imgs = preprocess_input(imgs)
    return imgs


# Load Real Images
real_datagen = ImageDataGenerator(rescale=1./255)
real_generator = real_datagen.flow_from_directory(
    subset_dir,
    target_size=(299, 299),  # Resize to 299x299 for InceptionV3
    batch_size=600,  # Load all 600 real images (100 per category)
    class_mode=None,
    shuffle=True
)
real_images = next(real_generator)

# Load Real Images
real_datagen = ImageDataGenerator(rescale=1./255)
real_generator = real_datagen.flow_from_directory(
    subset_dir,
    target_size=(299, 299),  # Resize to 299x299 for InceptionV3
    batch_size=600,  # Load all 600 real images (100 per category)
    class_mode=None,
    shuffle=True
)
real_images = next(real_generator)

Found 628 images belonging to 6 classes.


# Generate Images using VAE
def generate_images(model, latent_dim, num_images):
    z_sample = np.random.normal(size=(num_images, latent_dim))
    generated_images = model.predict(z_sample)
    return generated_images

generated_images = generate_images(decoder, latent_dim, 600)
generated_images_resized = tf.image.resize(generated_images, (299, 299))

# Generate Images using VAE
def generate_images(model, latent_dim, num_images):
    z_sample = np.random.normal(size=(num_images, latent_dim))
    generated_images = model.predict(z_sample)
    return generated_images

generated_images = generate_images(decoder, latent_dim, 600)
generated_images_resized = tf.image.resize(generated_images, (299, 299))

19/19 [==============================] - 6s 327ms/step


# Preprocess images for InceptionV3
real_images_preprocessed = preprocess_input(real_images)
generated_images_preprocessed = preprocess_input(generated_images_resized)

# Preprocess images for InceptionV3
real_images_preprocessed = preprocess_input(real_images)
generated_images_preprocessed = preprocess_input(generated_images_resized)


# Calculate FID
fid = calculate_fid(real_images_preprocessed, generated_images_preprocessed)
print('FID:', fid)

# Calculate FID
fid = calculate_fid(real_images_preprocessed, generated_images_preprocessed)
print('FID:', fid)

19/19 [==============================] - 155s 8s/step
19/19 [==============================] - 152s 8s/step
FID: 1.3319541321978917e+53


# Ensure generated images are in the correct format (299x299x3) for InceptionV3
generated_images_resized_inception = np.array(generated_images_resized)

# Ensure generated images are in the correct format (299x299x3) for InceptionV3
generated_images_resized_inception = np.array(generated_images_resized)


# Ensure images are preprocessed for InceptionV3 model
def preprocess_images(images):
    images = preprocess_input(images)
    return images

# Ensure images are preprocessed for InceptionV3 model
def preprocess_images(images):
    images = preprocess_input(images)
    return images


# Preprocess generated images
generated_images_preprocessed = preprocess_images(generated_images_resized_inception)

# Predict class probabilities for generated images
preds = inception_model.predict(generated_images_preprocessed)

# Preprocess generated images
generated_images_preprocessed = preprocess_images(generated_images_resized_inception)

# Predict class probabilities for generated images
preds = inception_model.predict(generated_images_preprocessed)

19/19 [==============================] - 158s 8s/step


def calculate_inception_score(preds, num_splits=10):
    scores = []
    split_size = preds.shape[0] // num_splits

    for i in range(num_splits):
        split_preds = preds[i * split_size:(i + 1) * split_size]
        p_yx = np.mean(split_preds, axis=0)  # Average class probabilities across the split
        scores.append(entropy(p_yx, base=2))  # Compute entropy of the average probabilities

    inception_score = np.exp(np.mean(scores))
    return inception_score

def calculate_inception_score(preds, num_splits=10):
    scores = []
    split_size = preds.shape[0] // num_splits

    for i in range(num_splits):
        split_preds = preds[i * split_size:(i + 1) * split_size]
        p_yx = np.mean(split_preds, axis=0)  # Average class probabilities across the split
        scores.append(entropy(p_yx, base=2))  # Compute entropy of the average probabilities

    inception_score = np.exp(np.mean(scores))
    return inception_score


# Calculate Inception Score
inception_score = calculate_inception_score(preds)
print('Inception Score:', inception_score)

# Calculate Inception Score
inception_score = calculate_inception_score(preds)
print('Inception Score:', inception_score)

Inception Score: 5902.863071669922


print('FID:', fid)

print('FID:', fid)

FID: 1.3319541321978917e+53


# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Generate one image for each category
def generate_images_per_category(model, latent_dim, num_images=1):
    generated_images = []
    for _ in range(num_images * len(categories)):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_images.append(generated_image[0])
    return np.array(generated_images)

# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Generate one image for each category
def generate_images_per_category(model, latent_dim, num_images=1):
    generated_images = []
    for _ in range(num_images * len(categories)):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_images.append(generated_image[0])
    return np.array(generated_images)


# Generate images
generated_images = generate_images_per_category(decoder, latent_dim, num_images=1)

# Generate images
generated_images = generate_images_per_category(decoder, latent_dim, num_images=1)

1/1 [==============================] - 0s 76ms/step
1/1 [==============================] - 0s 60ms/step
1/1 [==============================] - 0s 52ms/step
1/1 [==============================] - 0s 55ms/step
1/1 [==============================] - 0s 56ms/step
1/1 [==============================] - 0s 64ms/step


# Plot generated images
def plot_generated_images(images, categories):
    plt.figure(figsize=(20, 10))
    for i, img in enumerate(images):
        plt.subplot(1, len(categories), i + 1)
        plt.imshow((img * 255).astype(np.uint8))
        plt.axis('off')
        plt.title(categories[i])
    plt.show()

plot_generated_images(generated_images, categories)

# Plot generated images
def plot_generated_images(images, categories):
    plt.figure(figsize=(20, 10))
    for i, img in enumerate(images):
        plt.subplot(1, len(categories), i + 1)
        plt.imshow((img * 255).astype(np.uint8))
        plt.axis('off')
        plt.title(categories[i])
    plt.show()

plot_generated_images(generated_images, categories)


# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Function to generate and save faces
def save_generated_faces(model, latent_dim, categories, save_dir='/content/generated_faces'):
    os.makedirs(save_dir, exist_ok=True)
    for i, category in enumerate(categories):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_image = (generated_image[0] * 255).astype(np.uint8)
        plt.imsave(os.path.join(save_dir, f'{category}.png'), generated_image)
save_generated_faces(decoder, latent_dim, categories)

# Define the categories
categories = ['Angry', 'Surprise', 'Sad', 'Ahegao', 'Happy', 'Neutral']

# Function to generate and save faces
def save_generated_faces(model, latent_dim, categories, save_dir='/content/generated_faces'):
    os.makedirs(save_dir, exist_ok=True)
    for i, category in enumerate(categories):
        z_sample = np.random.normal(size=(1, latent_dim))
        generated_image = model.predict(z_sample)
        generated_image = (generated_image[0] * 255).astype(np.uint8)
        plt.imsave(os.path.join(save_dir, f'{category}.png'), generated_image)
save_generated_faces(decoder, latent_dim, categories)

1/1 [==============================] - 0s 59ms/step
1/1 [==============================] - 0s 51ms/step
1/1 [==============================] - 0s 59ms/step
1/1 [==============================] - 0s 52ms/step
1/1 [==============================] - 0s 50ms/step
1/1 [==============================] - 0s 61ms/step


import numpy as np
import tensorflow as tf
from tensorflow.keras import layers, models

# Example VAE definition
latent_dim = 2

# Encoder
encoder_inputs = tf.keras.Input(shape=(28, 28, 1))
x = layers.Flatten()(encoder_inputs)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim, name='z_mean')(x)
z_log_var = layers.Dense(latent_dim, name='z_log_var')(x)

# Sampling layer
def sampling(args):
    z_mean, z_log_var = args
    batch = tf.shape(z_mean)[0]
    dim = tf.shape(z_mean)[1]
    epsilon = tf.keras.backend.random_normal(shape=(batch, dim))
    return z_mean + tf.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling, output_shape=(latent_dim,), name='z')([z_mean, z_log_var])

encoder = models.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')

# Decoder
decoder_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(128, activation='relu')(decoder_inputs)
x = layers.Dense(28 * 28, activation='sigmoid')(x)
decoder_outputs = layers.Reshape((28, 28, 1))(x)

decoder = models.Model(decoder_inputs, decoder_outputs, name='decoder')

# VAE
outputs = decoder(encoder(encoder_inputs)[2])
vae = models.Model(encoder_inputs, outputs, name='vae')

# Compile and train the VAE
vae.compile(optimizer='adam', loss='binary_crossentropy')
(x_train, _), (x_test, _) = tf.keras.datasets.mnist.load_data()
x_train = np.expand_dims(x_train, -1) / 255.0
vae.fit(x_train, x_train, epochs=10, batch_size=32)

import numpy as np
import tensorflow as tf
from tensorflow.keras import layers, models

# Example VAE definition
latent_dim = 2

# Encoder
encoder_inputs = tf.keras.Input(shape=(28, 28, 1))
x = layers.Flatten()(encoder_inputs)
x = layers.Dense(128, activation='relu')(x)
z_mean = layers.Dense(latent_dim, name='z_mean')(x)
z_log_var = layers.Dense(latent_dim, name='z_log_var')(x)

# Sampling layer
def sampling(args):
    z_mean, z_log_var = args
    batch = tf.shape(z_mean)[0]
    dim = tf.shape(z_mean)[1]
    epsilon = tf.keras.backend.random_normal(shape=(batch, dim))
    return z_mean + tf.exp(0.5 * z_log_var) * epsilon

z = layers.Lambda(sampling, output_shape=(latent_dim,), name='z')([z_mean, z_log_var])

encoder = models.Model(encoder_inputs, [z_mean, z_log_var, z], name='encoder')

# Decoder
decoder_inputs = tf.keras.Input(shape=(latent_dim,))
x = layers.Dense(128, activation='relu')(decoder_inputs)
x = layers.Dense(28 * 28, activation='sigmoid')(x)
decoder_outputs = layers.Reshape((28, 28, 1))(x)

decoder = models.Model(decoder_inputs, decoder_outputs, name='decoder')

# VAE
outputs = decoder(encoder(encoder_inputs)[2])
vae = models.Model(encoder_inputs, outputs, name='vae')

# Compile and train the VAE
vae.compile(optimizer='adam', loss='binary_crossentropy')
(x_train, _), (x_test, _) = tf.keras.datasets.mnist.load_data()
x_train = np.expand_dims(x_train, -1) / 255.0
vae.fit(x_train, x_train, epochs=10, batch_size=32)

Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz
11490434/11490434 [==============================] - 0s 0us/step
Epoch 1/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.2306
Epoch 2/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.2022
Epoch 3/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1964
Epoch 4/10
1875/1875 [==============================] - 16s 9ms/step - loss: 0.1929
Epoch 5/10
1875/1875 [==============================] - 16s 9ms/step - loss: 0.1905
Epoch 6/10
1875/1875 [==============================] - 18s 10ms/step - loss: 0.1889
Epoch 7/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.1875
Epoch 8/10
1875/1875 [==============================] - 18s 9ms/step - loss: 0.1865
Epoch 9/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1856
Epoch 10/10
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1847

<keras.src.callbacks.History at 0x7869b3c3f460>


vae.save('vae_model.h5')

vae.save('vae_model.h5')

/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
  saving_api.save_model(


encoder.save('encoder_model.h5')
decoder.save('decoder_model.h5')

encoder.save('encoder_model.h5')
decoder.save('decoder_model.h5')

WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.
WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.


from google.colab import files

# Download the VAE model
files.download('vae_model.h5')

# Download the encoder and decoder models
files.download('encoder_model.h5')
files.download('decoder_model.h5')

from google.colab import files

# Download the VAE model
files.download('vae_model.h5')

# Download the encoder and decoder models
files.download('encoder_model.h5')
files.download('decoder_model.h5')


pip install streamlit

pip install streamlit

Collecting streamlit
  Downloading streamlit-1.36.0-py2.py3-none-any.whl (8.6 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 8.6/8.6 MB 13.1 MB/s eta 0:00:00
Requirement already satisfied: altair<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.2.2)
Requirement already satisfied: blinker<2,>=1.0.0 in /usr/lib/python3/dist-packages (from streamlit) (1.4)
Requirement already satisfied: cachetools<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (5.4.0)
Requirement already satisfied: click<9,>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.1.7)
Requirement already satisfied: numpy<3,>=1.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (1.25.2)
Requirement already satisfied: packaging<25,>=20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (24.1)
Requirement already satisfied: pandas<3,>=1.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.0.3)
Requirement already satisfied: pillow<11,>=7.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (9.4.0)
Requirement already satisfied: protobuf<6,>=3.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.20.3)
Requirement already satisfied: pyarrow>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (14.0.2)
Requirement already satisfied: requests<3,>=2.27 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.31.0)
Requirement already satisfied: rich<14,>=10.14.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (13.7.1)
Requirement already satisfied: tenacity<9,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.5.0)
Requirement already satisfied: toml<2,>=0.10.1 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.10.2)
Requirement already satisfied: typing-extensions<5,>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.12.2)
Collecting gitpython!=3.1.19,<4,>=3.0.7 (from streamlit)
  Downloading GitPython-3.1.43-py3-none-any.whl (207 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 207.3/207.3 kB 16.1 MB/s eta 0:00:00
Collecting pydeck<1,>=0.8.0b4 (from streamlit)
  Downloading pydeck-0.9.1-py2.py3-none-any.whl (6.9 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.9/6.9 MB 21.7 MB/s eta 0:00:00
Requirement already satisfied: tornado<7,>=6.0.3 in /usr/local/lib/python3.10/dist-packages (from streamlit) (6.3.3)
Collecting watchdog<5,>=2.1.5 (from streamlit)
  Downloading watchdog-4.0.1-py3-none-manylinux2014_x86_64.whl (83 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 83.0/83.0 kB 7.4 MB/s eta 0:00:00
Requirement already satisfied: entrypoints in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.4)
Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (3.1.4)
Requirement already satisfied: jsonschema>=3.0 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (4.19.2)
Requirement already satisfied: toolz in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.12.1)
Collecting gitdb<5,>=4.0.1 (from gitpython!=3.1.19,<4,>=3.0.7->streamlit)
  Downloading gitdb-4.0.11-py3-none-any.whl (62 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 62.7/62.7 kB 3.4 MB/s eta 0:00:00
Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2.8.2)
Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2023.4)
Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2024.1)
Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2.0.7)
Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2024.7.4)
Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (3.0.0)
Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (2.16.1)
Collecting smmap<6,>=3.0.1 (from gitdb<5,>=4.0.1->gitpython!=3.1.19,<4,>=3.0.7->streamlit)
  Downloading smmap-5.0.1-py3-none-any.whl (24 kB)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->altair<6,>=4.0->streamlit) (2.1.5)
Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (23.2.0)
Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (2023.12.1)
Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.35.1)
Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.19.0)
Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.10/dist-packages (from markdown-it-py>=2.2.0->rich<14,>=10.14.0->streamlit) (0.1.2)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.8.2->pandas<3,>=1.3.0->streamlit) (1.16.0)
Installing collected packages: watchdog, smmap, pydeck, gitdb, gitpython, streamlit
Successfully installed gitdb-4.0.11 gitpython-3.1.43 pydeck-0.9.1 smmap-5.0.1 streamlit-1.36.0 watchdog-4.0.1


import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io

import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io


# Load your trained VAE model
model = load_model('/content/vae_model.h5')

# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

# Load your trained VAE model
model = load_model('/content/vae_model.h5')

# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

2024-07-22 02:03:25.305 
  Warning: to view this Streamlit app on a browser, run it with the following
  command:

    streamlit run /usr/local/lib/python3.10/dist-packages/colab_kernel_launcher.py [ARGUMENTS]
2024-07-22 02:03:25.316 Session state does not function when running a script without `streamlit run`


!pip install streamlit numpy keras matplotlib

!pip install streamlit numpy keras matplotlib

Requirement already satisfied: streamlit in /usr/local/lib/python3.10/dist-packages (1.36.0)
Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (1.25.2)
Requirement already satisfied: keras in /usr/local/lib/python3.10/dist-packages (2.15.0)
Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (3.7.1)
Requirement already satisfied: altair<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.2.2)
Requirement already satisfied: blinker<2,>=1.0.0 in /usr/lib/python3/dist-packages (from streamlit) (1.4)
Requirement already satisfied: cachetools<6,>=4.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (5.4.0)
Requirement already satisfied: click<9,>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.1.7)
Requirement already satisfied: packaging<25,>=20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (24.1)
Requirement already satisfied: pandas<3,>=1.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.0.3)
Requirement already satisfied: pillow<11,>=7.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (9.4.0)
Requirement already satisfied: protobuf<6,>=3.20 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.20.3)
Requirement already satisfied: pyarrow>=7.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (14.0.2)
Requirement already satisfied: requests<3,>=2.27 in /usr/local/lib/python3.10/dist-packages (from streamlit) (2.31.0)
Requirement already satisfied: rich<14,>=10.14.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (13.7.1)
Requirement already satisfied: tenacity<9,>=8.1.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (8.5.0)
Requirement already satisfied: toml<2,>=0.10.1 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.10.2)
Requirement already satisfied: typing-extensions<5,>=4.3.0 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.12.2)
Requirement already satisfied: gitpython!=3.1.19,<4,>=3.0.7 in /usr/local/lib/python3.10/dist-packages (from streamlit) (3.1.43)
Requirement already satisfied: pydeck<1,>=0.8.0b4 in /usr/local/lib/python3.10/dist-packages (from streamlit) (0.9.1)
Requirement already satisfied: tornado<7,>=6.0.3 in /usr/local/lib/python3.10/dist-packages (from streamlit) (6.3.3)
Requirement already satisfied: watchdog<5,>=2.1.5 in /usr/local/lib/python3.10/dist-packages (from streamlit) (4.0.1)
Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (1.2.1)
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (0.12.1)
Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (4.53.1)
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (1.4.5)
Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (3.1.2)
Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib) (2.8.2)
Requirement already satisfied: entrypoints in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.4)
Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (3.1.4)
Requirement already satisfied: jsonschema>=3.0 in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (4.19.2)
Requirement already satisfied: toolz in /usr/local/lib/python3.10/dist-packages (from altair<6,>=4.0->streamlit) (0.12.1)
Requirement already satisfied: gitdb<5,>=4.0.1 in /usr/local/lib/python3.10/dist-packages (from gitpython!=3.1.19,<4,>=3.0.7->streamlit) (4.0.11)
Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2023.4)
Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas<3,>=1.3.0->streamlit) (2024.1)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)
Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2.0.7)
Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.27->streamlit) (2024.7.4)
Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (3.0.0)
Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/dist-packages (from rich<14,>=10.14.0->streamlit) (2.16.1)
Requirement already satisfied: smmap<6,>=3.0.1 in /usr/local/lib/python3.10/dist-packages (from gitdb<5,>=4.0.1->gitpython!=3.1.19,<4,>=3.0.7->streamlit) (5.0.1)
Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->altair<6,>=4.0->streamlit) (2.1.5)
Requirement already satisfied: attrs>=22.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (23.2.0)
Requirement already satisfied: jsonschema-specifications>=2023.03.6 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (2023.12.1)
Requirement already satisfied: referencing>=0.28.4 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.35.1)
Requirement already satisfied: rpds-py>=0.7.1 in /usr/local/lib/python3.10/dist-packages (from jsonschema>=3.0->altair<6,>=4.0->streamlit) (0.19.0)
Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.10/dist-packages (from markdown-it-py>=2.2.0->rich<14,>=10.14.0->streamlit) (0.1.2)


# Write the Streamlit app to a file
%%writefile app.py
import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io

# Write the Streamlit app to a file
%%writefile app.py
import streamlit as st
import numpy as np
from keras.models import load_model
import matplotlib.pyplot as plt
import io

Writing app.py


# Load your trained VAE model
model = load_model('/content/vae_model.h5')

# Load your trained VAE model
model = load_model('/content/vae_model.h5')


# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

    # Generate the image
    generated_image = model.predict(input_vector)

    # Scale the pixel values to [0, 255]
    generated_image = (generated_image * 127.5 + 127.5).astype(np.uint8)
    generated_image = generated_image[0]  # Remove the batch dimension

    # Display the image
    st.image(generated_image, use_column_width=True)

# Function to encode the keyword into a numerical format
def encode_keyword(keyword, vocab):
    encoded = np.zeros(len(vocab))
    if keyword in vocab:
        encoded[vocab.index(keyword)] = 1
    return encoded

# Example vocabulary
vocab = ['happy', 'sad', 'angry', 'surprised', 'neutral']

# Streamlit app
st.title("Image Generation using VAE")

keyword = st.selectbox('Select a keyword:', vocab)
generate_button = st.button('Generate Image')

if generate_button:
    condition = encode_keyword(keyword, vocab)

    # Generate random noise
    noise = np.random.normal(0, 1, (1, 100))  # Adjust the dimensions according to your model input

    # Combine noise and condition
    input_vector = np.concatenate([noise, condition.reshape(1, -1)], axis=1)

    # Generate the image
    generated_image = model.predict(input_vector)

    # Scale the pixel values to [0, 255]
    generated_image = (generated_image * 127.5 + 127.5).astype(np.uint8)
    generated_image = generated_image[0]  # Remove the batch dimension

    # Display the image
    st.image(generated_image, use_column_width=True)


!pip install pyngrok

!pip install pyngrok

Collecting pyngrok
  Downloading pyngrok-7.2.0-py3-none-any.whl (22 kB)
Requirement already satisfied: PyYAML>=5.1 in /usr/local/lib/python3.10/dist-packages (from pyngrok) (6.0.1)
Installing collected packages: pyngrok
Successfully installed pyngrok-7.2.0


from pyngrok import ngrok

# Run Streamlit in the background
!streamlit run app.py &

# Set up ngrok to tunnel to the Streamlit app
public_url = ngrok.connect(port='8501')
print(f"Streamlit app is live at: {public_url}")

from pyngrok import ngrok

# Run Streamlit in the background
!streamlit run app.py &

# Set up ngrok to tunnel to the Streamlit app
public_url = ngrok.connect(port='8501')
print(f"Streamlit app is live at: {public_url}")

Collecting usage statistics. To deactivate, set browser.gatherUsageStats to false.


  You can now view your Streamlit app in your browser.

  Local URL: http://localhost:8501
  Network URL: http://172.28.0.12:8501
  External URL: http://34.42.123.165:8501

Generated Image¶

Importing Libraries¶

Mounting the Google Drive¶

Reading the images from the drive¶

Displaying images form each folder¶

Load Images using "ImageDataGenerator" from keras.preprocessing.image¶

Creating a subset folder¶

Loading sample images¶

Loading the sample images¶

Building the Variational Autoencoder¶

Encoder¶

Decoder¶

Training¶

Model Evaluation¶

Generate Images¶

Generated Image¶

Importing Libraries¶

Mounting the Google Drive¶

Reading the images from the drive¶

Displaying images form each folder¶

Load Images using "ImageDataGenerator" from keras.preprocessing.image¶

Creating a subset folder¶

Loading sample images¶

Loading the sample images¶

Building the Variational Autoencoder¶

Encoder¶

Decoder¶

Training¶

Model Evaluation¶

Generate Images¶